java 通过 JNI 将 float[][] 传递给 C++ 的最简单方法
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/6751825/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
Simplest way to pass float[][] to C++ via JNI
提问by Haris Hasan
In my Java code I have a 2D float array float[x][4] floatArray. Here xcan be between 1 and 25. I have to pass this 2D float array to a C++method via JNI. My JNImethod is
在我的 Java 代码中，我有一个 2D 浮点数组float[x][4] floatArray。这里x可以在 1 到 25 之间。我必须通过 将这个二维浮点数组传递给一个C++方法JNI。我的JNI方法是
jboolean MyJNIMethod(JNIEnv * env, jobject obj, jobjectArray myArray)
{
//how to convert this myArray to something that can be safely passed to C++ method below
}
Inside MyJNIMethodI have to call a C++method and pass 2D float array taken from Java to this method
在内部MyJNIMethod我必须调用一个C++方法并将从 Java 获取的二维浮点数组传递给这个方法
bool MyCplusPlusMethod(float coordinates[][4])
{
}
I am having a hard time in properly converting jobject to float[][] due to lack of native development knowledge. Can anyone tell me the simplest and safest possible way? Thanks
由于缺乏本地开发知识，我很难将 jobject 正确转换为 float[][]。谁能告诉我最简单和最安全的方法？谢谢
回答by zacheusz
Something like this should work:
这样的事情应该工作：
jboolean MyJNIMethod(JNIEnv * env, jobject obj, jobjectArray myArray)
{
int len1 = env -> GetArrayLength(myArray);
jfloatArray dim= (jfloatArray)env->GetObjectArrayElement(myArray, 0);
int len2 = env -> GetArrayLength(dim);
float **localArray;
// allocate localArray using len1
localArray = new float*[len1];
for(int i=0; i<len1; ++i){
jfloatArray oneDim= (jfloatArray)env->GetObjectArrayElement(myArray, i);
jfloat *element=env->GetFloatArrayElements(oneDim, 0);
//allocate localArray[i] using len2
localArray[i] = new float[len2];
for(int j=0; j<len2; ++j) {
localArray[i][j]= element[j];
}
}
//TODO play with localArray, don't forget to release memory ;)
}
Note that this is outline. It won't compile ;) (I wrote it in this overstacks' editor)
请注意，这是大纲。它不会编译 ;) （我在这个 overstacks 的编辑器中写的）
In your class you should declare native method:
在您的课程中，您应该声明本机方法：
public native void myJNIMethod(float[][] m);
and in your c code corresponding:
并在您的 c 代码中对应：
JNIEXPORT jboolean JNICALL Java_ClassName_methodName (JNIEnv *, jobject, jobjectArray);
Here is JNI array operations documentation.
这是JNI 数组操作文档。
回答by Arpan Sarkar
For releasing the allocated memory you can do something like this:
要释放分配的内存，您可以执行以下操作：
static void releaseMatrixArray(JNIEnv *env, jobjectArray matrix) {
int size = (*env)->GetArrayLength(env, matrix);
for (int i = 0; i < size; i++) {
jfloatArray oneDim = (jfloatArray) (*env)->GetObjectArrayElement(env, matrix, i);
jfloat *elements = (*env)->GetFloatArrayElements(env, oneDim, 0);
(*env)->ReleaseFloatArrayElements(env, oneDim, elements, 0);
(*env)->DeleteLocalRef(env, oneDim);
}
}
release local Array reference:
释放本地数组参考：
free(localArray);
